EP2141868B1 - Service admission control - Google Patents

Abstract

The method involves determining whether a gateway (HGW1) e.g. home or residential gateway, connected to a terminal (T) and a connection point i.e. digital subscriber line access multiplexer (DSLAM) device, has required network resources e.g. transit time, for providing a service, from obtained information. The resources are reserved and reservation of the resources is informed to the gateway connected to the point for the provided service, if the gateway has the required network resources. An access to the service is rejected if the gateway does not have the required network resources. Independent claims are also included for the following: (1) a computer program comprising instructions for performing a method for processing a service accessing request (2) a gateway connected to a connection point, comprising an information obtaining unit.

Description

Background of the invention

The present invention relates to the general field of telecommunications.

It relates more particularly to the mechanisms for controlling access, by a terminal, to a service offered by a telecommunications network.

In general, such mechanisms consist in checking the availability of network resources before authorizing a service. They are also known as Call Admission Control (CAC) techniques. We can have an example in: US 2006/028981 .

In the current state of the art, there are admission control mechanisms distributed or distributed in a telecommunications network. Such mechanisms have been developed in particular within the framework of the IETF's RSVP (Resource ReSerVation Protocol) and Next Step In Signaling (NSIS) standards.

In these mechanisms, the terminal issues a request for access to a service in the form of dedicated signaling messages (also called signaling protocol) exchanged with the network. This signaling protocol follows the same path, in the telecommunications network, as the data associated with the service required by the user. In other words, it goes through the same network equipment as these data.

Each device verifies, in passing the signaling protocol, the local availability (that is to say at the level of the equipment strictly speaking) of the network resources necessary to provide the service, before transmitting the request to the equipment next. If, after passing through the various equipment concerned in the network, access to the service is authorized (availability of network resources from each equipment traversed), a positive acknowledgment message of the signaling protocol is sent to each equipment which, upon receipt of this message, proceeds to the configuration necessary to deliver the service (establishment of a context corresponding to the reservation of the resources, etc.).

Thus, in the distributed admission control mechanisms described above, the availability of end-to-end the entire telecommunications network with each equipment crossed by the data associated with the required service. These mechanisms therefore require that all of the network equipment traversed support the signaling protocol implemented, to be able to interpret the messages exchanged according to this protocol and to communicate according to this protocol.

Therefore, the application and deployment of such mechanisms is particularly complex and constraining in terms of implementation of the signaling.

Object and summary of the invention

• obtention d'informations relatives à des ressources réseau réservées par au moins une autre passerelle connectée à ce même point de raccordement, pour au moins un service fourni via ce point de raccordement ;
• vérification, à l'aide au moins de ces informations, qu'au moins la passerelle connectée au terminal et le point de raccordement disposent respectivement de ressources réseau nécessaires pour fournir le service requis par le terminal ;
• si tel est le cas :
  • ○ réservation de ces ressources réseau ; et
  • ○ information d'au moins une autre passerelle connectée au point de raccordement pour au moins un service fourni via ce point de raccordement de cette réservation ;
• sinon, rejet de l'accès au service.

The present invention makes it possible to overcome these disadvantages, by proposing, according to a first aspect, a method of processing a request for access to a service sent by a terminal, this method being implemented by a gateway connected to said terminal and to a connection point to a telecommunications network. According to the invention, the treatment method comprises the following steps:

• obtaining information relating to network resources reserved by at least one other gateway connected to this same connection point, for at least one service provided via this connection point;
• verification, using at least this information, at least the gateway connected to the terminal and the connection point respectively have network resources needed to provide the service required by the terminal;
• if that is the case :
  • ○ reservation of these network resources; and
  • ○ informing at least one other gateway connected to the connection point for at least one service provided via this connection point of this reservation;
• otherwise, rejection of access to the service.

• des moyens de réception d'une requête d'accès à un service émise par un terminal connecté à cette passerelle ;
• des moyens d'obtention d'informations relatives aux ressources réseau réservées par au moins une autre passerelle connectée à ce même point de raccordement pour au moins un service fourni via ce point de raccordement ;
• des moyens de vérification, à l'aide au moins de ces informations, qu'au moins la passerelle connectée au terminal et le point de raccordement disposent de ressources réseau nécessaires pour fournir le service requis par le terminal ;
• des moyens de réservation de ces ressources réseau et des moyens d'information de cette réservation d'au moins une autre passerelle connectée à ce même point de raccordement pour au moins un service fourni via ce point de raccordement, activés si tel est le cas ; et
• des moyens de rejet de l'accès au service activés sinon.

Correlatively, the invention also relates to a gateway connected to a connection point to a telecommunications network, and comprising:

• means for receiving a request for access to a service sent by a terminal connected to this gateway;
• means for obtaining information relating to the network resources reserved by at least one other gateway connected to this same connection point for at least one service provided via this connection point;
• verification means, using at least this information, at least the gateway connected to the terminal and the connection point have network resources necessary to provide the service required by the terminal;
• means for reserving these network resources and means for informing this reservation of at least one other gateway connected to this same connection point for at least one service provided via this connection point, activated if this is the case; and
• service access rejection means activated otherwise.

It should be noted that, throughout this document, gateway or access gateway is understood to mean a device interconnecting two communications networks, in particular, but not exclusively, a local area network (LAN). and a larger network of MAN (for Metropolitan Area Network) or WAN (for Wide Area Network). Such a local network may be, for example, a residential network (example of a residential or residential gateway), a corporate network, a network of a particular public place such as an airport, etc. By abuse of language, these various bridges are often described as residential bridges, even when they do not find use in the particular context of a dwelling.

Furthermore, the term connection point to a telecommunications network, any entity beyond the access gateway, in the telecommunications network, and to access the heart of the telecommunications network.

Thus, in a particular embodiment of the invention, the entity located at the boundary between the local loop (including in particular the copper line at the output of the access gateway) and the collection network (also called backhaul). The telecommunications network is a point of connection to the telecommunications network within the meaning of the invention. This entity is for example a DSLAM device (Digital Subscriber Line Access Multiplexer) for an xDSL type network, an Ethernet switch for a Gigabit-Ethernet type network, or an Optical Link Terminal (OLT) device for an optical type network based on on a PON (Passive Optical Network) architecture. By collection network is meant all the equipment located between this entity (for example the DSLAM) and the core network connection node, as well as the connection links between this entity and the core network (or backbone network).

Moreover, within the meaning of the invention, the term network resources of an entity of a telecommunications network, any type of resource (bandwidth in the upstream or downstream direction, memory, storage capacity of data packets, dimensioning of the data buses, etc.) made available by this entity to enable data exchange over a network. For example, the bandwidth offered at the output of an access gateway, on the link existing between the gateway and the connection point to the collection network, represents network resources of the access gateway in the sense of the invention. .

The invention therefore relates to a service access control function (also called call admission function) implemented in an access gateway to the telecommunications network (eg in a home or residential gateway). ) that takes into account the availability of network resources at the gateway and a network connection point.

According to the invention, it is therefore no longer necessary that all of the network equipment traversed support the signaling protocol implemented for a call admission control, since this call admission control is set implemented at the access gateway itself, and involves only the equipment located between the gateway and the connection point to the network in question.

In an advantageous embodiment of the invention, the connection point considered is a contention point identified in the telecommunications network, whatever it is.

Thus, in a particular context where contentions exist on the link between an access gateway and a connection point to the collection network, as for example on the link at the output of this connection point, and where the collection network is in itself properly dimensioned or even oversized in relation to the needs, it may be advantageous, in the context of the invention, to consider this point of connection to the collection network.

The invention thus supports the distribution of resources of the connection point between the different users connected to it, by limiting the contentions or the losses of packets.

In a particular embodiment of the invention in which the connection point in question is a connection point to the collection network of the telecommunications network, the invention advantageously makes it possible to manage the contentions that may exist in the local loop and, more generally, in the access network to the telecommunications network, by verifying the availability of resources at least at the access gateway (eg at the copper line) and at the connection point to the collection network (eg at the connection interface of the DSLAM to the collection network).

The access links respectively between the gateway and the point of connection to the collection network (network resource of the gateway in the sense of the invention) and between the connection point and the collection network (network resource of the connection point to the meaning of the invention) are today two major contention points in a telecommunications network. The collection network and the core network are generally oversized in relation to the needs of the services offered by the telecommunications network.

• d'une part, grâce à des informations obtenues par la passerelle connectée au terminal et relatives aux ressources réseau réservées par d'autres passerelles connectées au même point de raccordement que cette passerelle ; et
• d'autre part, grâce à la mise à disposition d'autres passerelles connectées au même point de raccordement que cette passerelle, des réservations qu'elle effectue.

The access control to a service required by a terminal is allowed, according to the invention:

• on the one hand, thanks to information obtained by the gateway connected to the terminal and relating to the network resources reserved by other gateways connected to the same connection point as this gateway; and
• on the other hand, through the provision of other gateways connected to the same connection point as this gateway, reservations it makes.

Thus, it is well understood that the performance of the access control mechanism according to the invention depends on the number of informed gateways and the number of gateways from which information is received in relation to the number of gateways connected to the connection point, and that it is possible to play on these numbers to reach an acceptable performance / complexity compromise as a function of the position of the connection point considered in the telecommunications network (ie the depth that separates this connection point from the core network).

Indeed, the complexity of implementation of the invention is limited by the number of subscribers (ie access gateways) present behind the same connection point.

Consequently, when the connection point in question is a point of connection to the collection network of the telecommunications network, such as a DSLAM, the deployment of the invention in the context of a large telecommunications network is facilitated, the number gateways behind such a connection point being relatively limited (for example, 1000 gateways). Thus the application of the invention, including in a large-sized network such as an xDSL (x Digital Subscriber Line) network, is permitted, ie, the scaling up of the access control mechanism proposed by the invention is possible.

This is all the more true that the processing method according to the invention can be easily deployed by an operator because it is implemented within the access gateway that the operator masters. It therefore does not require the addition of additional platform or network equipment updates (particularly in terms of signaling) other than the gateway unlike the techniques proposed in the prior art.

When the connection point to the telecommunications network in question is connected to a very large number of gateways, the use of partial information and the provision of resource reservation information to a limited number of these gateways can be envisaged. to limit the complexity of implementation of the invention.

In a particular embodiment of the invention, during the obtaining step, the information is received from at least one message broadcast to members of a multicast multicast group associated with the connection point and to which belongs the gateway connected to the terminal, and during the information step, said at least one other gateway is informed in a broadcast message to the members of this multicast broadcast group.

• à obtenir les informations à partir d'au moins un message diffusé aux membres dudit groupe de diffusion multicast ; et
• à informer ladite au moins une autre passerelle dans un message diffusé aux membres de ce groupe de diffusion multicast.

Correlatively, in this particular embodiment, the gateway according to the invention is configured to be a member of a multicast multicast group associated with the connection point and is adapted:

• obtaining the information from at least one broadcast message to members of said multicast broadcast group; and
• informing said at least one other gateway in a broadcast message to the members of this multicast broadcast group.

In this way, gateways members of the multicast broadcast group can easily communicate with each other and inform each other of the current bookings they manage.

Advantageously, the multicast broadcast group may be a multicast IP tree known to those skilled in the art.

The use of such a multicast IP broadcast tree is particularly advantageous when scaling up on large telecommunication networks, since it makes it possible to have to manage only one IP address for communications between the footbridges.

Alternatively, the multicast broadcast group may be an ATM or Ethernet multicast broadcast group.

• une bande passante disponible sur une interface physique en sortie du point de raccordement pour un ensemble de classes de service supportées par le point de raccordement ; et
• une bande passante disponible sur une interface physique en sortie du point de raccordement pour au moins une classe de service de cet ensemble.

By way of example, the information relating to the network resources obtained by the gateway can comprise in particular:

• a bandwidth available on a physical interface out of the connection point for a set of service classes supported by the connection point; and
• a bandwidth available on a physical interface out of the connection point for at least one class of service of this set.

For the purposes of the invention, the term "class of services" means a set of services having similar constraints in terms of quality of service, such as, for example, a class of delay-sensitive services, a class of loss-sensitive services, a class of jitter-sensitive services, etc.

In this way, the gateways according to the invention can perform an access control function based both on the availability of network resources on their link with the connection point and on the availability of network resources on the connection link of the network. connection point to the network.

• qu'une bande passante nécessaire pour délivrer le service requis par le terminal est disponible en sortie de la passerelle et en sortie du point de raccordement pour la classe de service associée au service demandé ;
• et qu'une bande passante nécessaire pour délivrer ce service est disponible sur l'interface physique en sortie de la passerelle et en sortie du point de raccordement.

They can check for example:

• that a bandwidth necessary to deliver the service required by the terminal is available at the gateway output and at the output of the connection point for the class of service associated with the requested service;
• and that a bandwidth necessary to deliver this service is available on the physical interface at the output of the gateway and at the output of the connection point.

The invention thus makes it possible to optimize the use of the resources at the gateway and the point of connection to the network, and to guarantee the quality of service to the applications with high added value by adapting the authorized bit rate for each class of service to level of the connection point.

By performing the two aforementioned tests at the access gateway and the connection point, it is ensured that each required service is treated identically, regardless of the class of service to which it belongs. The invention allows a fair access control: requests for access to non-priority services are no longer systematically refused, since a specific bit rate for these services (for the associated class of service) can be allocated and tested at the same time. gateway level. Service management within the same class of service and from one service class to another is optimized. The telecommunications network can thus be managed more efficiently and used better.

Alternatively, further checks can be made to determine if the necessary resources are sufficient at the connection point and the gateway. In particular, if the network connection point has within its architecture other contention levels than its connection link to the telecommunications network, they can be modeled so as to be verified during the verification step. Similarly, it is possible to transmit information about these contention levels in messages circulating on the broadcast group.

• une étape d'évaluation, à l'aide des informations :
  • ○ des ressources réseau réservées par plusieurs passerelles sur une interface physique en sortie du point de raccordement pour un ensemble de classes de service supportées par le point de raccordement ; et
  • ○ des ressources réseau réservées par plusieurs passerelles sur une interface physique en sortie du point de raccordement pour au moins une classe de service de cet ensemble ;
• une étape d'envoi d'au moins un message comprenant l'évaluation de ces ressources réseau réservées, à au moins une autre passerelle connectée au point de raccordement.

In a particular embodiment of the invention, the treatment method further comprises:

• an evaluation step, using the information:
  • ○ network resources reserved by multiple gateways on a physical interface out of the connection point for a set of service classes supported by the connection point; and
  • ○ network resources reserved by several gateways on a physical interface out of the connection point for at least one service class of this set;
• a step of sending at least one message comprising the evaluation of these reserved network resources to at least one other gateway connected to the connection point.

Advantageously, these steps can be implemented by a limited number of gateways connected to the connection point, for example by a so-called primary gateway, and this, so as to limit the complexity at the level of the other gateways.

• la réception :
  • ○ d'informations relatives aux ressources réseau réservées par plusieurs passerelles sur une interface physique en sortie du point de raccordement pour un ensemble de classes de service supportées par le point de raccordement ;
  • ○ d'informations relatives aux ressources réseau réservées par plusieurs passerelles sur une interface physique en sortie du point de raccordement pour au moins une classe de service dudit ensemble ;
• la mise à jour d'une base locale avec ces informations pour consultation lors de ladite vérification.

In a particular embodiment, the step of obtaining the information comprises:

• the reception :
  • ○ network resource information reserved by multiple gateways on a physical interface out of the connection point for a set of service classes supported by the connection point;
  • ○ information relating to the network resources reserved by several gateways on a physical interface out of the connection point for at least one service class of said set;
• updating a local database with this information for consultation during said verification.

The information received is, for example, the evaluations of the network resources sent by a primary gateway as mentioned above.

• la réception d'un message du point de raccordement comprenant :
  • ○ une bande passante disponible sur une interface physique en sortie du point de raccordement pour un ensemble de classes de service supportées par le point de raccordement ; et
  • ○ une bande passante disponible sur une interface physique en sortie du point de raccordement pour au moins une classe de service de cet ensemble ; et
• la mise à jour d'une base locale avec ces bandes passantes pour consultation lors de l'étape de vérification.

Alternatively, it may be information received from the connection point itself, if it is adapted to measure or evaluate this information. In particular, in a particular embodiment, the step of obtaining the information comprises:

• receiving a message from the connection point comprising:
  • ○ a bandwidth available on a physical interface at the end of the connection point for a set of service classes supported by the connection point; and
  • ○ a bandwidth available on a physical interface out of the connection point for at least one service class of this set; and
• updating a local database with these bandwidths for consultation during the verification step.

The receipt of these different information (information relating to the reserved network resources or available bandwidths) enables the gateways according to the invention to update a local database so that it can then proceed to the verification step of the method of treatment according to the invention.

The update can be of different natures. It may be to copy the information received in the local database of the gateway, because for example they are not evaluated directly by this gateway. As a variant, if the gateway itself evaluates this information, the information received may enable this gateway to verify that its own evaluation is correct, and the opposite case to rectify its evaluation. These two variants of realization will be considered as updates of a local base within the meaning of the invention.

In a particular embodiment, the processing method according to the invention further comprises a step of periodic sending by the gateway connected to the terminal of a message indicating the network resources it reserves for at least one service provided via the connection point to at least one other gateway connected to this connection point.

Thus, it is possible for the gateways receiving this periodic message to "re-synchronize" and to check the validity of their local copy of the evaluation of the resources reserved by the gateway.

This also allows these gateways to detect, in case of non receipt of this periodic message, phantom bookings associated with the gateway, that is to say, reservations released for example because of an unexpected disconnection of the gateway without that it has had time to inform the other gateways. The management of the resources of the telecommunications network is thus improved.

In a particular embodiment of the invention, the processing method further comprises, for at least one third-party equipment connecting the gateway connected to the terminal at the connection point, obtaining information relating to the resources reserved by at least one Another gateway connected to this equipment for at least one service provided via this equipment. In this embodiment, during the verification step, it is furthermore verified with the aid of this information, that the said at least one piece of equipment has the necessary network resources to provide the service, the step of reserving the resources being implemented if the gateway connected to the terminal, the connection point and said at least one equipment have the resources necessary to provide the service.

Correlatively, in this particular embodiment of the invention, the access gateway further comprises, for at least one device connecting the gateway connected to the terminal at the connection point, second means for obtaining information relating to the resources. reserved by at least one other gateway connected to this equipment for at least one service provided via this equipment. In addition, the means for verifying the gateway according to the invention also verify that the said at least one piece of equipment has the necessary network resources to provide the service, the reservation means being adapted to reserve these network resources if the gateway connected to the terminal, the connection point and said at least one device have the necessary network resources to provide the service required by the terminal.

This makes it possible to take into account other contention points identified in the network, for example at the level of an equipment comprised between the gateway and the connection point. This equipment can be in particular another DSLAM type equipment, in an architecture where several DSLAM are hierarchical, that is to say, cascaded before the point of connection to the collection network.

It may also be envisaged, in other embodiments, that the treatment method of the invention present in combination all or part of the aforementioned characteristics.

In a particular embodiment, the various steps of the method of processing a request according to the invention are determined by instructions of computer programs.

Consequently, the invention also relates to a computer program on an information carrier, this program being capable of being implemented in a gateway for access to a network or more generally in a computer, this program comprising instructions adapted to the implementation of the steps of a method of processing a request as described above.

This program can use any programming language, and be in the form of source code, object code, or intermediate code between source code and object code, such as in a partially compiled form, or in any other form desirable shape.

The invention also relates to a computer-readable information medium, comprising instructions of a computer program as mentioned above.

The information carrier may be any entity or device capable of storing the program. For example, the medium may comprise storage means, such as a ROM, for example a CD ROM or a microelectronic circuit ROM, or a magnetic recording medium, for example a diskette (floppy disc) or a disk hard or a rewritable recording means, for example a FLASH memory.

On the other hand, the information medium may be a transmissible medium such as an electrical or optical signal, which may be conveyed via an electrical or optical cable, by radio or by other means. The program according to the invention can be in particular downloaded on an Internet type network.

Alternatively, the information carrier may be an integrated circuit in which the program is incorporated, the circuit being adapted to execute or to be used in the execution of the method in question.

• un point de raccordement à un réseau de télécommunications ; et
• une pluralité de passerelles connectées au point de raccordement,

au moins une desdites passerelles étant conforme à l'invention telle que décrite précédemment pour traiter une requête d'accès audit au moins un service émise par un terminal connecté à cette passerelle.According to a second aspect, the invention also relates to an admission control system for at least one service, this system comprising at least:

• a point of connection to a telecommunications network; and
• a plurality of gateways connected to the connection point,

at least one of said gateways being in accordance with the invention as described above for processing an access request to said at least one service sent by a terminal connected to this gateway.

Thus, the invention proposes an admission control system for a service distributed over several access gateways, adapted to communicate with each other, for example via a multicast broadcasting group such as an IP multicast tree. . The admission control method is thus particularly robust and resistant to failures likely to be encountered by one or other of the gateways of the system according to the invention.

In a particular embodiment of the invention, at least one of the gateways is connected to the connection point via at least one other device.

Brief description of the drawings

• la figure 1 représente un système de contrôle d'admission conforme à l'invention dans son environnement, dans un mode particulier de réalisation ;
• la figure 2 représente sous forme schématique, une passerelle d'accès conforme à l'invention dans un mode particulier de réalisation ;
• la figure 3 représente sous forme d'organigrammes, les principales étapes mises en oeuvre au cours du procédé de traitement d'une requête selon l'invention, lorsqu'il est mis en oeuvre par une passerelle d'accès conforme à l'invention du système de contrôle d'admission représenté sur la figure 1 ;
• la figure 4 représente de façon schématique les principaux états d'une passerelle conforme à l'invention dans un second mode de réalisation ;
• la figure 5 représente un système de contrôle d'admission conforme à l'invention dans son environnement, dans un troisième mode de réalisation de l'invention.

Other features and advantages of the present invention will emerge from the description given below, with reference to the accompanying drawings which illustrate an embodiment having no limiting character. In the figures:

• the figure 1 represents an admission control system according to the invention in its environment, in a particular embodiment;
• the figure 2 represents in schematic form, an access gateway according to the invention in a particular embodiment;
• the figure 3 represents in the form of flowcharts, the main steps implemented during the method of processing a request according to the invention, when it is implemented by an access gateway according to the invention of the control system admission shown on the figure 1 ;
• the figure 4 schematically shows the main states of a gateway according to the invention in a second embodiment;
• the figure 5 represents an admission control system according to the invention in its environment, in a third embodiment of the invention.

Detailed description of several embodiments

In the various examples envisaged here, one places oneself in the context of an xDSL type telecommunications network based on an Asynchronous Transfer Mode (Asynchronous Transfer Mode) or Gigabit Ethernet type of collection network architecture.

However, the invention also applies to other types of telecommunications networks, such as for example optical telecommunications networks based on a Gigabit Passive Optical Network (GPON) architecture, for managing network resources at the gateway level. access, ONT (Optical Network Terminal) equipment and OLT (Optical Link Terminal) equipment.

Moreover, in the various embodiments described here, it is checked, before authorizing access to a service, the availability of the bandwidth on the output link of various equipment (network resources of these equipment in the sense of the 'invention). The bandwidth considered can be understood in the upstream direction and / or in the downward direction. In the remainder of the description, the term bandwidth is what is offered by a physical link and throughput that is consumed by one or more services (or applications). In a manner known per se, a bit rate and a bandwidth are expressed in the same unit namely in bit / s (or multiples thereof, for example in kbit / s or in Mbps).

These hypotheses are however in no way limiting, the invention being applicable also to the control of other network resources (in particular in addition to the bandwidth), such as for example a loss rate, a transit delay, a variation of transit delay, etc. In particular, it is possible to control, in addition to the bandwidth, the internal contention at the connection point, such as for example a shared bus.

In addition, in the various embodiments described here, is considered as a connection point, a connection point to the collection network of the telecommunications network. Of course, the invention applies also to other points of connection to the telecommunications network, such as a point of connection to the core network of the telecommunications network.

The figure 1 represents a CAC-S admission control system according to the invention in a first embodiment, and implemented in an xDSL type NW telecommunications network. The invention is of course also applicable to other types of telecommunication networks, as indicated previously in this document.

According to the invention, the admission control system CAC-S comprises a plurality of access gateways HGW _i , i = 1, .., n, according to the invention. These bridges are for example here residential or domestic gateways (or "homegateway" in English).

In the embodiment described here, each gateway HGW _i according to the invention has the hardware architecture of a computer as shown in FIG. figure 2 .

It comprises in particular a processor 1, a random access memory type RAM 2, a read-only memory ROM 3, a rewritable non-volatile flash memory type 4, means 5 of communication with the network NW, and means 6 of communication with T terminals (for example a WiFi or LAN interface).

ROM-type read-only memory 3 constitutes a recording medium in accordance with the invention on which is recorded a computer program according to the invention adapted to execute the main steps of the method of processing a request according to the invention. represented in the form of an organization chart figure 3 described later.

The gateways HGW _{i provide} the interface between the telecommunications network NW and one or more terminals T, such as for example computers, telephones, etc.

They are connected to the collection network RC of the telecommunications network NW by a DSLAM device (connection point within the meaning of the invention), equipment of the telecommunications network NW known per se, and whose main function is to recover the data streams. transiting over the telephone lines to which it is connected and to multiplex this data to direct them to the collection network.

It is assumed here that the DSLAM device supports a predetermined number K of classes of service offered by the network of NW telecommunications. These classes of service are, for example, the classes of service specified by the IETF or the ITU. They are known to those skilled in the art and will not be detailed further here.

In the first embodiment described here, the DSLAM device is configured with a multicast IP address rated @ IP , dedicated to the admission control function according to the invention. It provides, using this address, a multicast IP tree @MC to which the HGW gateways _i , i = 1, .., n are subscribed. The set of gateways subscribed to the multicast @MC tree constitutes a multicast group associated with the DSLAM device within the meaning of the invention.

In a manner known to those skilled in the art, such a multicast distribution or distribution tree enables each member of a group of subscribers to this tree to transmit and / or receive messages via a group IP address. multicast (each group member uses this multicast IP address as long as they do not leave the group). Further details on the operation and implementation of multicast IP trees (assignment of multicast IP addresses, subscription management, multicast tree construction, etc.) are available in particular in RFC 966, 1112, 1301, 1458, 2189, 2236, 2627, 2908, 3171, 3228 and 3376 of the IETF.

The multicast tree @MC here has the particularity of having only one replication point, namely the DSLAM device itself, and n sheets, namely the n domestic gateways HGW _i . It is advantageously configured here to allow each home gateway HGW _i to send and receive multicast messages. Thus configured, the multicast @MC tree behaves like a bus on which each member of the group of subscribers to this tree can send a message to all members already connected.

We will now describe, with reference to the figure 3 , the main steps of the method of processing a request according to the invention, when they are implemented by a home gateway HGW _i according to the invention and belonging to the admission control system shown in FIG. figure 1 . Subsequently, the labels of the signaling messages exchanged on the broadcast tree are given for illustrative purposes only.

When connecting the HGW gateway _i to the DSLAM device (step E10), the latter is configured with the multicast @ IP IP address of the multicast @MC tree, by the operator management system (not shown on the figure). This management system is for example as described in document TR-69 of the DSL Forum.

• les valeurs nominales de bande passante, notées ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}$
  
  (j), j=1,..,K, allouées au niveau du dispositif DSLAM à chaque classe de service j supportée par ce dispositif ;
• la valeur nominale de bande passante offerte par l'interface physique de raccordement du dispositif DSLAM au réseau de collecte RC (i.e. bande passante nominale disponible en sortie du dispositif DSLAM), notée ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{PHY}};$
  
  et
• les valeurs nominales de bande passante, notées ${\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{CoS}}$
  
  (j), (j), j=1,..,K, allouées au niveau de la passerelle HGW_i à chaque classe de service j.

During this configuration step, it also receives, here from the operator management system:

• nominal bandwidth values, rated ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}$
  
  ( j ), j = 1, .., K, allocated at the level of the DSLAM device to each class of service j supported by this device;
• the bandwidth nominal value offered by the DSLAM device's physical connection interface to the RC collection network (ie nominal bandwidth available at the DSLAM device output), noted ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{PHY}};$
  
  and
• nominal bandwidth values, rated ${\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{CoS}}$
  
  ( j ), ( j ), j = 1, .., K, allocated at the level of the gateway HGW _i to each class of service j .

Furthermore, during this step, the HGW gateway _i also accesses the nominal value of the bandwidth offered by the physical connection interface of the HGW gateway _i to the DSLAM device (ie nominal bandwidth available at the output of the HGW gateway _i ), rated ${\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{PHY}}\mathrm{.}$

In a manner known to those skilled in the art, this value depends on the access gateway considered, and can vary in particular as a function of the synchronization rate of the ADSL interface of the gateway with the DSLAM device at the time of its connection. This rate depends, among other things, on the distance of the gateway with the DSLAM device. Access to this value is known per se and will not be detailed further here. It is, in general, accessible via the firmware that manages the component implementing the ADSL interface in the gateway.

(j) et ${\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{CoS}}$

(j), j=1,..,K, peuvent être exprimées en termes de pourcentages des valeurs de bande passante ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{PHY}}$

et ${\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{PHY}}$

respectivement. Selon la politique d'allocation retenue en matière de bandes passantes (ou de débits) et de services, ces pourcentages peuvent être recouvrant sur l'ensemble des classes de service, c'est-à-dire que la somme des pourcentages alloués aux différentes classes de service peut être supérieure à 100%.It should be noted that the bandwidth values considered for the classes of service, namely the values ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}$

( j ) and ${\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{CoS}}$

( j ), j = 1, .., K, can be expressed in terms of percentages of the bandwidth values ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{PHY}}$

and ${\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{PHY}}$

respectively. According to the allocation policy adopted for bandwidth (or bandwidth) and services, these percentages can be recovered over all classes of service, ie the sum of the percentages allocated to the different types of service. classes of service can be greater than 100%.

The gateway HGW _i then subscribes to the multicast tree @MC associated with the DSLAM device using the multicast @ IP address obtained during its configuration (step E20), for example using the IGMP protocol (Internet Group Management Protocol). The skilled person is invited to refer to RFC 3376 cited above, for details on the steps implemented by an entity to subscribe to an IP multicast tree.

Once subscribed to the multicast tree @MC, the gateway HGW _i sends a HELLO connection message intended in particular to make itself known to other gateways HGW _m , m = 1, .., n, m ≠ i, connected to the device DSLAM and subscribed to the @MC tree.

This HELLO message makes it possible, in particular, for each gateway HGW _m , m = 1,..., N, m ≠ i subscribed to the multicast tree @MC, to update a local database comprising a list of the gateways connected to the DSLAM device and subscribed to the multicast @MC tree. In particular, upon receipt of this HELLO message, the HGW gateways _m create in their respective bases, a new context for the gateway HGW _i . Here by context of the HGW gateway _i , a set of information associated with this gateway and including the network resources it reserves, as detailed later.

Following the transmission of the HELLO message, the HGW gateway _i receives via the multicast @MC tree (step E30), information representative of the state of the current network resource reservations for each of these other gateways, for at least a service provided via the DSLAM device.

This information is received for example in SYNCHRO messages sent separately by each gateway HGW _m , m ≠ i, on the @MC tree, following the reception of the HELLO message. Each SYNCHRO message contains here the rate reserved by the gateway at the origin of this message for each class of service j , j = 1, .., K.

Note Λ _CoS ( m , j ) the rate reserved by the gateway HGW _m , for service requests associated with the service class j . This rate also characterizes the bit rate reserved by the HGW gateway _m at the DSLAM device (and in particular on its connection link to the collection network), for services associated with the service class j.

où ${\mathrm{\Lambda }}_{\mathit{req}}^j\left(m,p\right)$

désigne le débit associé à la réservation p gérée par la passerelle HGW_m et correspondant à un service de classe j.It is defined by: ${\mathrm{\Lambda }}_{\mathit{CoS}}\left(m,j\right)={\displaystyle \underset{p}{\mathrm{\Sigma }}}{\mathrm{\Lambda }}_{\mathit{req}}^j\left(m,p\right)$

or ${\mathrm{\Lambda }}_{\mathit{req}}^j\left(m,p\right)$

denotes the rate associated with the reservation p managed by the gateway HGW _m and corresponding to a class service j .

On receiving a SYNCHRO message from a gateway HGW _m , the gateway HGW _i updates a local database DB, stored in its memory. rewritable volatile memory 4, creating for this gateway HGW _m a context comprising the information Λ _cos ( m, j ).

Note that, at the start of the gateway HGW _i , the values Λ _Cos ( i, j ) of the gateway HGW _i for j = 1, .. K, are initialized to zero.

• du débit total qu'elle réserve pour l'ensemble des services, noté Λ _PHY (i) : $${\mathrm{\Lambda }}_{\mathit{PHY}}\left(i\right)=\mathrm{\sum }{\mathrm{\Lambda }}_{\mathit{CoS}}\left(i,j\right)$$
  
• du débit total réservé pour chaque classe de service par l'ensemble des passerelles HGW_m, m=1, ..n, et noté Λ _CoS (j), j=1,..,K ; et
• du débit total réservé par l'ensemble des passerelles HGW_m, m=1, ..n, connectées au dispositif DSLAM, noté Λ _PHY .

In addition, the HGW gateway _i also stores in the DB database representative local variables:

• the total throughput that it reserves for all services, noted Λ _PHY ( i ): $${\mathrm{\Lambda }}_{\mathit{PHY}}\left(i\right)=\mathrm{\Sigma }{\mathrm{\Lambda }}_{\mathit{CoS}}\left(i,j\right)$$
  
• the total rate reserved for each class of service by the set of gateways HGW _m , m = 1, ..n, and denoted by Λ _CoS ( j ), j = 1, .., K; and
• the total flow reserved by all the gateways HGW _m , m = 1, ..n, connected to the device DSLAM, noted Λ _PHY .

Thus, following the reception and storage of the values Λ _Cos ( m , j ), for m = 1, .., n, and j = 1, .. K, the gateway HGW _i updates the local variables Λ _PHY and Λ _Cos ( j ), ( j ), j = 1, .., K, for example using the following equations: $${\mathrm{\Lambda }}_{\mathit{CoS}}\left(j\right)={\displaystyle \underset{m=1}{\overset{\mathrm{not}}{\mathrm{\Sigma }}}}{\mathrm{\Lambda }}_{\mathit{CoS}}\left(m,j\right)\mathrm{and}{\mathrm{\Lambda }}_{\mathit{PHY}}={\displaystyle \underset{j=1}{\overset{K}{\mathrm{\Sigma }}}}{\mathrm{\Lambda }}_{\mathit{CoS}}\left(j\right)$$

• réception d'une requête d'accès RA d'un terminal T lié à la passerelle ;
• libération de ressources réservées par la passerelle ;
• réception d'un message sur l'arbre multicast @MC relatif à l'état des réservations des ressources réseau réservées par les autres passerelles.

The gateway HGW _i then goes on standby (step E40) of at least one of these events:

• receiving an access request RA from a terminal T linked to the gateway;
• release of resources reserved by the bridge;
• receiving a message on the multicast tree @MC relating to the state of reservation of the network resources reserved by the other gateways.

On receipt of an access request RA issued by a terminal T connected to the gateway HGW _i (step E50), the gateway HGW _i consults its database DB (step E60) to check whether the network resources necessary to provide this service are available, on the one hand at its own level, that is to say here on its connection link to the DSLAM device, and on the other hand, at the level of the DSLAM device, that is to say here on the connection link of the DSLAM device to the RC collection network.

It is assumed that the requested service in the RA request belongs to the class of service and there is Λ c _req flow necessary to provide that service.

• disponibilité des ressources au niveau de la passerelle HGW_i : $${\mathrm{\Lambda }}_{\mathit{req}}+{\mathrm{\Lambda }}_{\mathit{CoS}}\left(i,c\right)\le {\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{CoS}}\left(c\right)$$
  
  $${\mathrm{\Lambda }}_{\mathit{req}}+{\mathrm{\Lambda }}_{\mathit{PHY}}\left(i\right)\le {\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{PHY}}$$
  
• disponibilité des ressources au niveau du dispositif DSLAM : $${\mathrm{\Lambda }}_{\mathit{req}}+{\mathrm{\Lambda }}_{\mathit{CoS}}\left(c\right)\le {\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}\left(c\right)$$
  
  $${\mathrm{\Lambda }}_{\mathit{req}}+{\mathrm{\Lambda }}_{\mathit{PHY}}\le {\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{PHY}}$$
  

The gateway HGW _i checks in particular (step E70) if the following four conditions are met, using the information stored in its database DB:

• resource availability at the HGW gateway _i : $${\mathrm{\Lambda }}_{\mathit{req}}+{\mathrm{\Lambda }}_{\mathit{CoS}}\left(i,\mathrm{vs}\right)\le {\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{CoS}}\left(\mathrm{vs}\right)$$
  
  $${\mathrm{\Lambda }}_{\mathit{req}}+{\mathrm{\Lambda }}_{\mathit{PHY}}\left(i\right)\le {\mathrm{\Lambda }}_{\mathrm{HGWi}}^{\mathrm{PHY}}$$
  
• availability of resources at the DSLAM device level: $${\mathrm{\Lambda }}_{\mathit{req}}+{\mathrm{\Lambda }}_{\mathit{CoS}}\left(\mathrm{vs}\right)\le {\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}\left(\mathrm{vs}\right)$$
  
  $${\mathrm{\Lambda }}_{\mathit{req}}+{\mathrm{\Lambda }}_{\mathit{PHY}}\le {\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{PHY}}$$
  

If at least one of these conditions is not verified, then the home access gateway HGW _i denies access to the service by the terminal T (step E80).

If all conditions are verified, then the HGW gateway _i reserves the network resources necessary to provide the service required by the terminal T (step E90).

$${\mathrm{\Lambda }}_{\mathit{PHY}}\left(i\right)\leftarrow {\mathrm{\Lambda }}_{\mathit{PHY}}\left(i\right)+{\mathrm{\Lambda }}_{\mathit{req}}$$

$${\mathrm{\Lambda }}_{\mathit{CoS}}\left(c\right)\leftarrow {\mathrm{\Lambda }}_{\mathit{CoS}}\left(c\right)+{\mathrm{\Lambda }}_{\mathit{req}}$$

$${\mathrm{\Lambda }}_{\mathit{PHY}}\leftarrow {\mathrm{\Lambda }}_{\mathit{PHY}}+{\mathrm{\Lambda }}_{\mathit{req}}$$

In addition, it updates, in its database, the variables representative of this reservation, with the flows consumed by this reservation, that is to say: $${\mathrm{\Lambda }}_{\mathit{CoS}}\left(i,\mathrm{vs}\right)\leftarrow {\mathrm{\Lambda }}_{\mathit{CoS}}\left(i,\mathrm{vs}\right)+{\mathrm{\Lambda }}_{\mathit{req}}$$

$${\mathrm{\Lambda }}_{\mathit{PHY}}\left(i\right)\leftarrow {\mathrm{\Lambda }}_{\mathit{PHY}}\left(i\right)+{\mathrm{\Lambda }}_{\mathit{req}}$$

$${\mathrm{\Lambda }}_{\mathit{CoS}}\left(\mathrm{vs}\right)\leftarrow {\mathrm{\Lambda }}_{\mathit{CoS}}\left(\mathrm{vs}\right)+{\mathrm{\Lambda }}_{\mathit{req}}$$

$${\mathrm{\Lambda }}_{\mathit{PHY}}\leftarrow {\mathrm{\Lambda }}_{\mathit{PHY}}+{\mathrm{\Lambda }}_{\mathit{req}}$$

The other variables remain unchanged.

The gateway HGW _i then informs the other gateways of this reservation by sending on the multicast @MC tree a message containing the updated bit rate values that it consumes, namely A _CoS ( i , j ), for j = 1 , ... K and A _PHY ( i ) (step E100).

This message is for example an UPDATE or SYNCHRO message. It is received simultaneously by all access gateways connected to the DSLAM device and subscribed to the multicast @TM broadcast tree, which can then update their own local databases with these values. It also contains, in this first embodiment, the local values updated by the gateway HGW _i variables Λ _CoS ( j ) for j = 1, ... K and Λ _PHY .

Once the service is delivered to the terminal T by the telecommunications network NW, the network resources allocated to this service are released by the gateway HGW _i (step E110).

The gateway HGW _i then updates its database DB (that is, the service belonging to the service class c, the variables Λ _CoS ( i , c ), Λ _PHY ( i ), Λ _CoS ( c ) and Λ _PHY by subtracting the value Λ _req ) (step E120).

It then sends a message on the multicast tree @MC, for example a RELEASE message, in order to inform the other gateways of the availability of these network resources previously reserved for the service. This message comprises, in the first embodiment described here, the updated values of the variables _CoS ( i , c ), _PHY ( i ), _CoS ( c ) and _PHY .

Upon receipt of an UPDATE or SYNCHRO message transmitted on the multicast @MC tree by another HGW gateway _x (x ≠ i) connected to the DSLAM device, indicating a reservation of network resources for a service associated with a class of service by this gateway (step E130), and including including (new) rates Λ _CoS ( x , s ) and Λ _PHY ( x ) consumed by the gateway HGW _x , the gateway HGW _i updates its database ( step E140) with these values.

Furthermore, as mentioned above, in the first embodiment described here, each UPDATE or SYNCHRO message furthermore comprises the local values maintained by the HGW gateway _x , the variables _ΛCoS ( s ) and _ΛPHY . Thus, upon receipt of such a message, the gateway HGW _i also checks the new values of the local variables Λ _CoS ( s ) and Λ _PHY according to, for example, the following equations (update in the sense of the invention): $${\mathrm{\Lambda }}_{\mathit{CoS}}\left(s\right)\equiv {\displaystyle \underset{m=1}{\overset{\mathrm{not}}{\mathrm{\Sigma }}}}{\mathrm{\Lambda }}_{\mathit{CoS}}\left(m,s\right)\mathrm{and}{\mathrm{\Lambda }}_{\mathit{PHY}}\equiv {\displaystyle \underset{m=1}{\overset{\mathrm{not}}{\mathrm{\Sigma }}}}{\mathrm{\Lambda }}_{\mathit{PHY}}\left(m\right)$$

Similarly, upon receipt of a RELEASE message transmitted on the multicast @MC tree by another HGW gateway _x (x ≠ i) connected to the DSLAM device, indicating a release of network resources for a service associated with a class of service s by this gateway (step E130), and including in particular the new rates Λ _CoS ( x , s ) and Λ _PHY ( x ) consumed by the gateway HGW _x , the gateway HGW _i updates its database (step E140) with these values.

Moreover, as previously described, in the first embodiment described here, each RELEASE message further comprises the local values, maintained by the HGW gateway _x , variables Λ _CoS ( s ) and Λ _PHY . Thus, upon receipt of such a message, the gateway HGW _i also checks the new values of the local variables Λ _CoS ( s ) and Λ _PHY according to, for example, the following equations (update in the sense of the invention): $${\mathrm{\Lambda }}_{\mathit{CoS}}\left(s\right)\equiv {\displaystyle \underset{m=1}{\overset{\mathrm{not}}{\mathrm{\Sigma }}}}{\mathrm{\Lambda }}_{\mathit{CoS}}\left(m,s\right)\mathrm{and}{\mathrm{\Lambda }}_{\mathit{PHY}}\equiv {\displaystyle \underset{m=1}{\overset{\mathrm{not}}{\mathrm{\Sigma }}}}{\mathrm{\Lambda }}_{\mathit{PHY}}\left(m\right)$$

When a gateway no longer consumes any resources, it sends an UPDATE or SYNCHRO message to the multicast @MC tree with the rate values set to 0 and then stops issuing UPDATE or SYNCHRO messages.

Each HGW gateway _i also sends, at regular intervals (corresponding to a period T _ref ), on the multicast tree @MC, an UPDATE or SYNCHRO message to inform the other gateways of the amount of network resources that it uses for management. of his current bookings.

This is intended to allow each gateway receiving this message to (re) synchronize easily, and in particular to verify that the HGW gateway _i uses the resources reported in the local database of the gateway (update in the sense of the invention).

This mechanism thus makes it possible to avoid "ghost" reservations, ie bookings reported in the local database of a gateway but which no longer exist in practice, and which would eventually consume the entire band. passing without the possibility of releasing it. This mechanism is known to those skilled in the art under the term "Graceful state release" and thus implements a monitoring of reservations in case of problems, such as the sudden electric shutdown of a gateway HGW _i without it having had the material time to issue a RELEASE message.

All other gateways HGW _m (m ≠ i) then perform a release of the resources consumed by the gateway HGW _i (similar to the receipt of an implicit RELEASE message), if they have not received an UPDATE or SYNCHRO message at the end of a time T _ref .

The broadcast of such a message at regular intervals also allows a gateway coming to connect to the DSLAM device to update its local database (see step E30).

It should also be noted that the reception of a HELLO message sent by a gateway that has just connected to the DSLAM device may also make it possible to combat "ghost" reservations.

• crée un nouveau contexte dans sa base de données locale pour la passerelle HGW_i, si aucun contexte associé à la passerelle HGW_i n'existe dans la base ; ou
• réinitialise (i.e. met à zéros) les variables locales de débit associées à la passerelle HGW_i, si un contexte existe pour la passerelle HGW_i dans la base locale.

Indeed, on receiving a HELLO message sent by an HGW gateway _i , a gateway HGW _m :

• creates a new context in its local database for the HGW gateway _i , if no context associated with the HGW gateway _i exists in the base; or
• resets (ie sets to zeros) the local rate variables associated with the HGW gateway _i , if a context exists for the HGW gateway _i in the local database.

Thus, in the event of a sudden stop of the gateway HGW _i without it having had time to send a RELEASE message, the other gateways can update their local base of reconnection of the gateway HGW _i to the device DSLAM and send a HELLO message after subscription to the multicast @MC tree.

In an alternative embodiment, in order to manage the collisions that may be encountered during the reservation of the resources, when all the conditions are verified in step E70, the gateway HGW _i sends on the multicast tree @MC a REQUEST message containing his reservation. All other gateways receiving this message will then update their respective local databases.

If, after a time equal to 2 * T _m (predetermined T _m ), necessary for the propagation of the REQUEST reservation request of the gateway HGW _i and its taking into account by the other gateways, the gateway HGW _i receives no reservation request from another gateway, then the HGW gateway _i confirms its reservation by issuing a CONFIRM message with the rate values it consumes, namely with the values Λ _CoS ( i , j ) and Λ _PHY ( i ), and also, in the first embodiment described here, the values of the variables Λ _CoS ( j ) and Λ _PHY for j = 1, .. K.

If, however, two or more gateways simultaneously issue a REQUEST message, then a collision will be detected by the HGW gateway _i before a time equal to 2 * T _m . If the resources are not sufficient to provide these requests simultaneously, then in this variant, all requests are canceled. Each gateway can retransmit its request after a random time.

The time T _m can be chosen so as to represent the maximum propagation time of the messages on the multicast tree, that is to say: $${\mathit{T}}_{\mathit{m}}\mathit{=}\mathit{tPU}\mathit{+}\mathit{Tpd}\mathit{+}\mathit{Tr}$$

Tpu representing the propagation time in the upstream direction, Tpd the downstream propagation delay and Tr the replication time of the IP multicast packet in the DSLAM device. As an indication, on an xDSL link, Tpu = Tpd = Tp >> Tr , ie T _m = 2 * Tp ≈160ms.

In a second embodiment of the invention, among the access gateways HGW _m , m = 1, n, connected to the DSLAM device and described above, a so-called primary gateway and possibly a so-called gateway are also considered. secondary. The selection process of these gateways can be based in particular on the MAC or IP address of the access gateways. For example, the gateway with the smallest or largest MAC address is selected as the primary gateway.

The operation of the other gateways is similar to that described previously in the first embodiment.

The primary gateway has the additional role (in addition to implementing the method of processing a request according to the invention), to issue at regular intervals a synchronization message SYNCHRO (for example at each time interval T _ref multiple of T _m ). This message contains, in the embodiment described here, the updated values of the variables Λ _CoS ( j ) and Λ _PHY respectively reflecting the total rate reserved by the gateways HGW _i , i = 1, .., n connected to the device DSLAM and the total rate reserved by the gateways HGW _i , i = 1, .., n for each service j , j = 1, .., K.

Upon receiving these values, each gateway re-synchronizes itself, i.e updates and / or checks its local database.

respectivement aux valeurs nominales ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}$

(j) reçues par la passerelle primaire au moment de sa configuration par le système de gestion de l'opérateur. Du point de vue d'une passerelle conforme à l'invention, les valeurs de débit disponibles dans chaque classe de service et sur l'interface physique en sortie du dispositif DSLAM donnent une indication de la bande passante réservée par les différentes passerelles connectées au dispositif DSLAM $({\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{phy}}$

et ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}$

(j) étant connues de chaque passerelle). Par conséquent, il s'agit également à proprement parler d'informations représentatives des ressources réseau réservées par des passerelles connectées au dispositif DSLAM au sens de l'invention. De façon équivalente, chaque passerelle peut stocker dans sa base de données ces valeurs en remplacement (ou en sus) des valeurs Λ _CoS (j) et Λ _PHY et s'en servir à l'étape E70, moyennant l'adaptation des équations (Eq.1) à (Eq.4), pour vérifier qu'une bande passante nécessaire pour fournir le service requis par le terminal est disponible en sortie de la passerelle et en sortie du dispositif DSLAM pour une classe de service associée au service requis par le terminal et qu'une bande passante nécessaire pour fournir ce service est disponible sur l'interface physique en sortie de la passerelle et en sortie du dispositif DSLAM.In an alternative embodiment, the primary gateway sends, in the SYNCHRO message, the bandwidth values available in each service class j , j = 1, .. K, as well as the bandwidth value available on the physical interface. at the output of the DSLAM device. These values are calculated by the primary gateway by subtracting the value Λ _PHY , respectively the values Λ _CoS ( j ), at the nominal value ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{PHY}},$

respectively at nominal values ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}$

( j ) received by the Primary Gateway at the time of its configuration by the Operator Management System. From the point of view of a gateway according to the invention, the bit rate values available in each class of service and on the physical interface at the output of the DSLAM device give an indication of the bandwidth reserved by the different gateways connected to the device. DSLAM $({\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{phy}}$

and ${\mathrm{\Lambda }}_{\mathrm{DSLAM}}^{\mathrm{CoS}}$

( j ) being known from each gateway). Consequently, it is also, strictly speaking, information representative of the network resources reserved by gateways connected to the DSLAM device within the meaning of the invention. Equivalently, each gateway can store in its database these values replacing (or in addition) the values Λ _CoS ( j ) and Λ _PHY and use them in step E70, by means of adapting the equations (Eq.1) to (Eq.4) to verify that a bandwidth necessary to provide the service required by the terminal is available at the gateway output and at the output of the DSLAM device for a class of service associated with the service required by the terminal and that a bandwidth necessary to provide this service is available on the physical interface at the output of the gateway and at the output of the DSLAM device.

The secondary gateway monitors the primary gateway and if it disappears (detected for example because no synchronization frame SYNCHRO has been detected for a time equal to 2 * T _ref ), it becomes primary gateway and triggers the process election of a new secondary bridge.

Furthermore, on receiving a CONFIRM, UPDATE or RELEASE message sent by a gateway HGW _m on the multicast tree @MC, the primary gateway stores the consumption in terms of the bandwidth of the gateway HGW _m and updates its base. local data as previously described. If a CONFIRM or UPDATE message is not received before a T _ref interval of a gateway that has issued a REQUEST reservation message, the primary gateway invalidates the reservation. The RELEASE message is not confirmed, but the primary gateway can validate the new state by sending a new SYNCHRO message.

In addition, if the primary gateway detects that a gateway is no longer consuming resources, the primary gateway removes it from the list of gateways to monitor.

The secondary gateway performs the same calculations and updates as the primary gateway so that it can take over the primary gateway instantly when needed.

Note that in this second embodiment of the invention, only the primary gateway (and possibly the secondary gateway) is adapted to send SYNCHRO messages. Furthermore, in this second embodiment, it is not necessary that the gateways include in the messages RELEASE, CONFIRM and UPDATE the values of the variables _ΛCoS ( j ) and Λ _PHY from their local database, these values being evaluated and sent at regular time intervals by the primary gateway, in the SYNCHRO message.

The figure 4 summarizes the main signaling messages exchanged between the gateways and the main states that can be taken by a gateway HGW _i of the system 1 according to the invention in this second embodiment.

The START state refers to the state of the HGW gateway _i when connected to the DSLAM device. The HELLO message is the message sent by the gateway after subscribing to the multicast @MC tree associated with the DSLAM device.

The SYNC state refers to a synchronization state in which the HGW gateway _i is located when receiving SYNCHRO or RELEASE frames.

A SYNCHRO message is a message emitted at regular intervals T _ref by the primary or secondary gateway in order to synchronize all the gateways and to allow them to update their local databases.

This message contains the total bit rate reserved by the HGW gateways _m connected to the DSLAM device, and the total bit rate reserved by these gateways for each service.

As a variant, it may contain the bandwidths available at the output of the DSLAM device per class of service as well as the total bandwidth available at the output of the DSLAM device.

In yet another variant (see first embodiment), when there are no primary and secondary gateways, this message may also contain the total bit rate reserved by the gateway at the origin of the message as well as the rate reserved by this gateway for each class of service.

If the gateway HGW _i has reservations in progress (NbResa ≠ 0), in this case it passes from the SYNC state to a RESERVED state, and sends at regular intervals T _ref an UPDATE message on the multicast @TM tree containing the the flows it reserves for each class of service and the total rate reserved.

After sending a REQUEST message on the multicast @MC tree to inform the other gateways of a desired reservation by the HGW gateway _i (the REQUEST message contains the desired bit rate and the class of service), this one goes into a WAIT standby state and activates a counter using a clock not shown on the figure 1 .

If no REQUEST message is received by the gateway during a time interval equal to 2 * T _m , then the gateway confirms its previously requested reservation with a CONFIRM message, and enters a RESERVED state.

Otherwise, the gateway HGW _i goes into a COLLISION state. If enough resources are available at the DSLAM device for support the simultaneous requests issued, then the gateway HGW _i goes into a RESERVED state, otherwise in a SYNC state.

• On peut considérer une variante dans laquelle il n'y a pas de passerelles primaire et secondaire, ni de message SYNCHRO. Les messages UPDATE servent toujours à resynchroniser l'ensemble des passerelles connectées au dispositif DSLAM mais chaque passerelle insère en plus de sa consommation la vue locale des ressources du dispositif DSLAM. La re-synchronisation s'effectue localement selon le même procédé que celui décrit précédemment. Chaque passerelle surveille ainsi les autres.
• Dans une autre variante, les messages CONFIRM et REQUEST sont remplacés par un message UPDATE annonçant la nouvelle consommation de la passerelle. Un message SYNCHRO peut être envoyé pour confirmer ou infirmer la réservation, et/ou pour confirmer ou infirmer une libération des ressources soit par la passerelle primaire soit par la passerelle qui a émis la requête initiale ou libéré les ressources.
• Dans une autre variante encore, le dispositif DSLAM mesure en permanence le débit utilisé dans chaque classe de service sur le lien de raccordement vers le réseau de collecte BB. A cette fin, le dispositif DSLAM se base notamment sur l'état de ses buffers sur l'interface de sortie ou sur une mesure effectuée sur son interface de sortie et accessible via son système de gestion, telle que par exemple le nombre de paquets ou d'octets écoulés depuis 5 minutes ou 30 secondes. Il diffuse alors à intervalles réguliers (par exemple T_ref), sur l'arbre multicast @MC en direction des passerelles, cette information dans un message SYNCHRO. Ainsi, dans cette variante, le dispositif DSLAM joue le rôle de la passerelle primaire. Les passerelles se servent de l'information contenue dans le message SYNCHRO pour effectuer l'admission d'appel et/ou mettre à jour la bande passante restante dans chaque classe de service supportée par le DSLAM.

The states and messages considered in the embodiments described above represent exemplary embodiments and are in no way limiting. Other variants could be envisaged. For example :

• One can consider a variant in which there is no primary and secondary gateways, nor message SYNCHRO. The UPDATE messages are always used to resynchronize all the gateways connected to the DSLAM device, but each gateway inserts in addition to its consumption the local view of the resources of the DSLAM device. Re-synchronization is carried out locally according to the same method as that described previously. Each gateway thus monitors others.
• In another variant, the CONFIRM and REQUEST messages are replaced by an UPDATE message announcing the new consumption of the gateway. A SYNCHRO message may be sent to confirm or deny the reservation, and / or to confirm or deny a release of resources either by the primary gateway or by the gateway that issued the initial request or released the resources.
• In yet another variant, the DSLAM device continuously measures the bit rate used in each class of service on the connection link to the collection network BB. To this end, the DSLAM device is based in particular on the state of its buffers on the output interface or on a measurement performed on its output interface and accessible via its management system, such as for example the number of packets or bytes passed for 5 minutes or 30 seconds. It then broadcasts at regular intervals (for example T _ref ), on the multicast tree @MC towards the gateways, this information in a SYNCHRO message. Thus, in this variant, the DSLAM device plays the role of the primary gateway. The gateways use the information contained in the SYNCHRO message to perform call admission and / or update the remaining bandwidth in each DSL supported class of service.

We will now describe, with reference to the figure 5 , a third embodiment of the invention, in which other points of contention of the local loop that the access gateway and the connection point are controlled by the access gateway according to the invention before allowing access to a service by a terminal T.

In this third embodiment, it is assumed that at least one gateway among the gateways of the CAC-S system, namely here the gateways HGW ₁ , HGW ₂ and HGW ₃ , are connected to the device DSLAM via less another equipment.

In the example described here, only one other device is considered, which is for example a second DSLAM device denoted DSLAM '. The device DSLAM 'is configured with a multicast IP address rated @ IP ', dedicated to the admission control function according to the invention. It provides, with the help of this address, a multicast IP tree @MC 'to which the gateways HGW ₁ , HGW ₂ and HGW ₃ are subscribed.

Thus, in this example, the method according to the invention makes it possible to manage the copper line (link between the gateway and the DSLAM device), the link to the collection network (link between the DSLAM device and the RC collection network). ), and the link between DSLAM and DSLAM devices'.

As a variant, it is possible to consider a plurality of devices between the gateway and the point of connection to the collection network RC (and the contention points associated with them), such as, for example, several DSLAM devices connected in cascade before the point of connection to the collection network (hierarchy of several DSLAMs), each equipment being configured with its own multicast IP address and providing the gateways connected to it with a multicast tree.

The operation of the gateways HGW ₄ , ..., HGW _n and the DSLAM device is identical to that described in the two previous embodiments.

The operation of the gateways HGW ₁ , HGW ₂ and HGW ₃ is also similar, except that these gateways transmit information messages of their current reservations and receive messages relating to the resources reserved by the other gateways, on the one hand on the multicast tree @MC and secondly on the multicast tree @MC '.

• qu'une bande passante nécessaire pour fournir ce service est disponible en sortie de la passerelle, en sortie du dispositif DSLAM' et en sortie du dispositif DSLAM, pour la classe de service associée à ce service ;
• qu'une bande passante nécessaire pour fournir ce service est disponible sur l'interface physique en sortie de la passerelle, en sortie du dispositif DSLAM' et en sortie du dispositif DSLAM.

Moreover, during the step E70, they check, for a given service required by the terminal:

• that a bandwidth necessary to provide this service is available at the output of the gateway, at the output of the DSLAM device and at the output of the DSLAM device, for the class of service associated with this service;
• that a bandwidth necessary to provide this service is available on the physical interface at the output of the gateway, at the output of the DSLAM device and at the output of the DSLAM device.

où ${\mathrm{\Lambda }}_{\mathrm{g}}^{\mathrm{CoS}}\left(c\right)$

et ${\mathrm{\Lambda }}_{\mathrm{g}}^{\mathrm{PHY}}$

désignent respectivement le débit alloué à la classe de service c sur le lien de niveau g contrôlé et le débit physique du lien de niveau g contrôlé, ${\mathrm{\Lambda }}_{\mathit{req}}^c\left(m,j\right)$

désigne le débit associé à la réservation j gérée par la passerelle HGW_m et correspondant à un service de classe cet Λ _req (m, j) désigne le débit associé à la réservation j gérée par la passerelle HGW_m (cette réservation j correspondant à une classe de service quelconque parmi les classes de service supportées).Thus, in general, if one considers CG points of congestion (or contention) between a gateway HGW _i and the point of connection to the collection network (CG≥2, and in the example of the figure 5 , CG = 3), and that the set G (g), for g = 1, .., CG, denotes the gateways contributing to the congestion of the link of level g, a service required by a terminal linked to the gateway and belonging to a service class c will be accepted by the gateway if we check for the CG congestion points in which the gateway participates, the two following inequalities (using the notations previously introduced): $${\displaystyle \underset{m\in \mathrm{BOY\; WUT}\left(\mathrm{boy\; Wut}\right)}{\mathrm{\Sigma }}}{\displaystyle \underset{j}{\mathrm{\Sigma }}}{\mathrm{\Lambda }}_{\mathit{req}}^{\mathrm{vs}}\left(m,j\right)\le {\mathrm{\Lambda }}_{\mathrm{boy\; Wut}}^{\mathrm{CoS}}\left(\mathrm{vs}\right)\mathrm{and}{\displaystyle \underset{m\in \mathrm{BOY\; WUT}\left(\mathrm{boy\; Wut}\right)}{\mathrm{\Sigma }}}{\displaystyle \underset{j}{\mathrm{\Sigma }}}{\mathrm{\Lambda }}_{\mathit{req}}\left(m,j\right)\le {\mathrm{\Lambda }}_{\mathrm{boy\; Wut}}^{\mathrm{PHY}}$$

or ${\mathrm{\Lambda }}_{\mathrm{boy\; Wut}}^{\mathrm{CoS}}\left(\mathrm{vs}\right)$

and ${\mathrm{\Lambda }}_{\mathrm{boy\; Wut}}^{\mathrm{PHY}}$

denote respectively the bit rate allocated to the service class c on the controlled level link g and the physical bit rate of the controlled level link g , ${\mathrm{\Lambda }}_{\mathit{req}}^{\mathrm{vs}}\left(m,j\right)$

refers to the rate associated with the reservation j managed by the HGW gateway _m and corresponding to a service class that Λ _req (m, j) is the rate associated with the reservation j managed by the HGW _m gateway (this reservation j corresponding to a any class of service among the service classes supported).

In the various embodiments described above, it is assumed that all the HGW gateways _i connected to the DSLAM device are configured using the IP address associated with this multicast tree and are subscribed to this tree. It is the same in the third embodiment for the device DSLAM '. Thus, the admission control function implemented by a gateway HGW _i according to the invention takes into account the reservations managed by the set of gateways HGW _i connected to the device DSLAM (and possibly to the device DSLAM 'in the third embodiment).

This hypothesis is however not limiting. Indeed, in an alternative embodiment corresponding to a suboptimal operating mode with respect to the previously described embodiments, only a subset of gateways among the gateways connected to the DSLAM device (respectively DSLAM ') are configured with the @ IP address associated with the multicast IP tree @MC (respectively with the @ IP address associated with the IP multicast @TM tree) and subscribed to this tree. Thus, in this variant, only the reservations managed by the subset of gateways subscribed multicast trees will be taken into account when implementing the method of processing a request according to the invention.

Similarly, in the embodiments described here, all the classes of service supported by the connection point considered to the telecommunications network (eg DSLAM device) are taken into account. As a variant, only a subset of this set of service classes can be considered when implementing the method of processing a request according to the invention.

• la diffusion des chaines de télévision en multicast hiérarchique (toutes les chaines du bouquet ne sont pas acheminées aux DSLAM, seules les plus regardées le sont, les autres chaines l'étant en fonction des demandes et des ressources) ;
• la vidéo à la demande, de façon similaire ;
• la voix sur IP afin d'optimiser le remplissage du conduit « voix » en sortie du DSLAM tout en offrant une excellente qualité de service.

The invention thus makes it possible to manage all the contentions that may exist between the gateways and the connection link to the collection network of a telecommunications network, namely in particular the contentions on the copper line between the gateway and the DSLAM and the contentions on the DSLAM connection link to the collection network). This solution can be advantageously applied to various services such as:

• the broadcasting of television channels in hierarchical multicast (all channels of the bouquet are not routed to DSLAM, only the most watched are, the other channels being according to requests and resources);
• video on demand, similarly;
• voice over IP to optimize the filling of the voice channel at the output of the DSLAM while offering an excellent quality of service.

Claims (14)

1. Method for processing a request to access a service sent by a terminal (T), said method being implemented by a gateway (HGW₁) connected to the terminal and to a point of connection (DSLAM) to a telecommunication network (NW),
   said processing method being characterized in that it comprises the following steps:

   - obtaining (E30, E60, E130) information relating to network resources reserved by at least one other gateway connected to this same point of connection, for at least one service supplied via this point of connection;

   - verifying (E70), using at least this information, that at the least the gateway connected to the terminal and the point of connection have network resources necessary to supply the service;

   - if such is the case:

   o reserving (E90) these network resources; and

   o communicating (E100) this reservation for at least one other gateway connected to the point of connection for at least one service supplied via this point of connection;

   - otherwise, refusing (E80) of access to the service.
2. Method for processing a request according to Claim 1, characterized in that, during said obtaining step (E130), said information is received from at least one message broadcast to the members of a multicast broadcast group associated with the point of connection and to which said gateway belongs,
   and in that, during said communication step (E100), said at least one other gateway is informed in a message broadcast to said members of said multicast broadcast group.
3. Method for processing a request according to Claim 1, characterized in that it includes:

   - a step of evaluating, using said information:

   ○ the network resources reserved by a number of gateways on a physical interface at the output of the point of connection for a set of classes of service supported by the point of connection; and

   o the network resources reserved by a number of gateways on a physical interface at the output of the point of connection for at least one class of service of said set;

   - a step of sending at least one message comprising the evaluation of these reserved network resources, to at least one other gateway connected to said point of connection.
4. Processing method according to Claim 3, characterized in that said sending step is implemented periodically, according to a determined period.
5. Method for processing a request according to Claim 1, characterized in that obtaining said information comprises:

   - receiving:

   ○ information relating to the network resources reserved by a number of gateways on a physical interface at the output of the point of connection for a set of classes of service supported by the point of connection;

   ○ information relating to the network resources reserved by a number of gateways on a physical interface at the output of the point of connection for at least one class of service of said set; and

   - updating a local database with this information for consultation during said verification.
6. Method for processing a request according to Claim 1, characterized in that it also includes a step of periodically sending, by said gateway connected to said terminal, a message indicating the network resources that it reserves for at least one service supplied via the point of connection to at least one other gateway connected to said point of connection.
7. Method for processing a request according to Claim 1, characterized in that the information relating to the network resources comprises:

   - a bandwidth available on a physical interface at the output of the point of connection for a set of classes of service supported by the point of connection; and

   - a bandwidth on a physical interface at the output of the point of connection for at least one class of service of said set.
8. Method for processing a request according to Claim 7, characterized in that obtaining said information comprises:

   - receiving a message from the point of connection including said bandwidth available for a set of service classes and said bandwidth available for at least one class of service; and

   - updating a local database with these bandwidths for consultation during said verification.
9. Method for processing a request according to Claim 1, characterized in that it also comprises, for at least one third-party element (DSLAM') linking the gateway to the point of connection, the obtaining of information relating to the resources reserved by at least one other gateway connected to this element for at least one service supplied via this element, and in that, during the verification step (E70), this information is also used to check that said at least one element has the network resources necessary to supply the service requested by the terminal, the step of reservation (E90) of the resources being implemented if the gateway connected to the terminal, the point of connection and said at least one element have said resources necessary to supply the service.
10. Computer program comprising instructions for executing the steps of the method for processing a request according to Claim 1 when said program is run by a computer.
11. Gateway (HGW₁) connected to a point of connection (DSLAM) to a telecommunication network (NW), said gateway comprising means for receiving a request to access a service sent by a terminal connected to this gateway,
    said gateway being characterized in that it also comprises:

    - means for obtaining information relating to the network resources reserved by at least one other gateway connected to this same point of connection for at least one service supplied via this point of connection;

    - means for verifying, using at least this information, that at least the gateway connected to the terminal and the point of connection have network resources necessary to supply the service requested by the terminal;

    - means for reserving these network resources and means for communicating this reservation to at least one other gateway connected to this same point of connection for at least one service supplied via this point of connection, activated if such is the case; and

    - means for refusing access to the service, activated otherwise.
12. Gateway according to Claim 11, characterized in that it is configured to be a member of a multicast broadcast group (@MC) associated with the point of connection and in that it is adapted:

    - to obtain said information from at least one message broadcast to the members of said multicast broadcast group; and

    - to inform said at least one other gateway in a message broadcast to said members of said multicast broadcast group.
13. Gateway according to Claim 12, characterized in that said multicast broadcast group is a multicast IP tree (@MC).
14. Admission control system (CAC-S) controlling admission to at least one service, characterized in that it comprises:

    - a point of connection (DSLAM) to a telecommunication network (NW);
    and

    - a plurality of gateways (HGW₁) connected to the point of connection, at least one of said gateways conforming to Claim 11 in order to process a request to access said at least one service sent by a terminal connected to this gateway.

Images